Not a big deal, but I wonder why this dispersement needs to take place. Searches become more difficult, and imprecise.

Logically, we should index everything first by year then by date. But years ago we decided not to for technical reasons. The thought was that since less than 2% of the games have the date-info filled in anyhow, it would make the index vastly larger/slower for not much of a benefit. And so we have indexes such as "ECO,year" and "Player,year".

When you sort something by year and there are a number of games with the same years, the order that they get spit out is effectively random. It's not truly random: it really depends on all sorts of behind-the-scenes details including deleted records and other boring things. Only MySQL knows for sure. For our purposes, it's random.

Since those early days, that 2% figure is now closer to 10%, and with the new Admin-Game Editor we'll be getting even more. The argument can be made that now we finally should index everything by year and date, and not just year. We agree. However that's a big change, with possible side-effects we haven't considered, so we've been slow in attempting to implement it.

I remember a few years ago having a discussion, or a lecture given even, by a member named <klangenfarben>, who lamented the PGN indices and proposed a new and improved system vastly superior to the PGN. That ship seems to have sailed. Stimulating discussion nonetheless.

Perhaps that should read "reached". A spot check showed that games from the last few years, especially tournament games, include the full date field. That is probably because the date is readily available when the PGN is created.

Going back to previous years, even recently submitted games often lack the date field not because of carelessness, but because the information is simply not readily available. Many older games have been found in newspaper columns, magazines, or books which would give only the year.

To illustrate: on page 174 the chapter reads as follows, "Kissingen, 12-24 August, 1928" - and indeed dates with the individual games are not, or coincidentally given.

Meticulously organized and well-written book, but the Games Register at the end only gives City and Year, not date.

I sorta like it; that's why we have to do the research. The book I have in front of me was published in <1949> ... "Databases? PGNs? What??!!"

[Event "London Match"]
[Site "London, England"]
[Date "1866.??.??"]
[EventDate "?"]
[Round "6"]
[Result "0-1"]
[White "Bird, Henry E"]
[Black "Steinitz, Wilhelm"]
[ECO "A02"]
[WhiteElo "?"]
[BlackElo "?"]

1.f4 e5 2.fxe5 d6 3.exd6 Bxd6 4.Nf3 Nf6
5.d4 Nc6 6.Bg5 Bg4 7.e3 Qd7 8.Bxf6 gxf6
9.Bb5 O-O-O 10.d5 Qe7 11.Bxc6 Qxe3+ 12.Qe2 Qc1+ 13.Qd1 Rde8+ 14.Bxe8 Rxe8+ 15.Kf2 Qe3+ 16.Kf1 Bxf3 17.gxf3 Bc5 18.Kg2 Rg8+ 0-1

Each bit of information is labelled by a specific tag or code; for example, <[White ""]>. If you do not use this format, the computer cannot tell properly interpret the information. Your error message "PGN lacks the White code" was probably the result of not having the <[White""]> code. It's also possible that you erased the quotation marks (""). I do that sometimes too, and they are essential.

Here are six Webb-Hartston games:

We quite often receive heroic efforts of chess fans who tediously typed every move by hand, even converting from descriptive to algebraic. Then when we process the game we find that a game of 80 moves might have two or three typos. Sometimes we can fix them by hand, other times they are too mysterious to fix, either way it's a pain for all involved.